2026/09/06

Sunday, September 6, 2026 (#249)

Saturday Sunday Monday ⇒ next

References 2026/09/05 (← links) 2026/09/12 (← links) 2026/09/13 (← links)

Main Work

a careful re-think of how Ferreteria connection-sessions are structured

A conundrum I keep getting myself into, when I want to have an object available directly and also as a Q-var:

• If the truth-source is the Q-var, there's no way to trigger hooks on the actual object when it's set. (This applies even when there's no direct access.)
• If the truth-source is the direct property, there's no (reliable) way to make the direct property settable from the Q-var.
  • This is what the Source and Proxy clades were trying to solve.

Solution: A Q-Proxy which supports hooks... or something...?

• How about: Q-var which expects to be set up with an eventable object.
• On get and set hooks, it calls Get and Set events in the eventable.
• The eventable handles the actual getting and setting.

...but this only works well if there's only one truth-source to update. If we have more than one, then the event-objects need to include a lookup-key... or become the source-of-truth, which probably misses the point completely.